Scientists Develop Artificial Synaptic Device Mimicking Brain to Revolutionize Information Technology

In their breakthrough, scientists at INST utilized two-dimensional electron gas (2DEG) within the EuO-KTaO₃ oxide heterostructure to create a neuromorphic chip.

 

Researchers at the Institute of Nano Science and Technology (INST), Mohali, have made a groundbreaking leap towards transforming information technology by developing an artificial synaptic device that mimics the behaviour of biological synapses, offering the potential for more efficient and faster computing models. This development brings neuromorphic architecture, inspired by the human brain, closer to reality, promising to overcome the limitations of traditional von Neumann computing.

Conventional computing systems, based on the von Neumann architecture, separate memory and computation into independent physical units, leading to a bottleneck when handling large volumes of data and complex operations. This separation slows down processing as all tasks must travel through a single channel to memory and back. In contrast, the human brain conducts computational operations and memory access directly, offering reconfigurable and dynamic processing. Inspired by this, neuromorphic electronics aim to create more efficient systems that can learn and adapt, similar to neurons in the brain.

In their breakthrough, scientists at INST utilized two-dimensional electron gas (2DEG) within the EuO-KTaO₃ oxide heterostructure to create a neuromorphic chip. This chip emulates cognitive functions, including learning, memory, and sensory perception, while demonstrating resistive switching behaviour. The device showed the ability to conduct logic gate operations and replicated short- and long-term synaptic plasticity, key properties of biological synapses.

A key feature of the device is its persistent photoconductivity, where light shining on the EuO-KTaO₃ interface generates a current that continues even after the light is turned off. This property mirrors how biological synapses retain information, making the device particularly suitable for cognitive and neuromorphic computing applications.

The research, funded by India's Department of Science and Technology (DST) under its Nano mission, as well as CSIR, has resulted in an advanced combinatorial pulsed laser deposition setup that enabled the development of this chip. Prof. Suvankar Chakraverty, who led the project, highlighted the potential of neuromorphic design in oxide interfaces for more energy-efficient and faster processing, especially in artificial intelligence (AI) applications and device miniaturization.

The study was published in the journal Applied Physics Letters.

The implications of this work are vast, with possible applications in sectors like healthcare, education, and environmental sustainability. Neuromorphic systems offer resilience, fault tolerance, and the ability to learn and change over time, paving the way for more personalized and responsive technologies that improve quality of life.